Crash, reboot and WHEA 18 Cache Hierarchy Error after loading up The Sims 4.

[emothxughts]

(PC specs in the drop down under my profile picture. Post copied from Linus Tech Tips forums, no reply yet at the time this was posted.)

Previously, when I crashed, rebooted and WHEA 18'd in games, I was running a Gigabyte RX5600XT card and I narrowed the problem down to the card, as detailed in a thread of mine on LTT forums and this thread on TPU forums. So I sent the card out for repair and put in the card I'm currently running now, a Sapphire Pulse RX550 2GB card. It has been running flawlessly in games like Genshin Impact, Skyrim, Destiny 2, Kingdom Come Deliverance so far, if somewhat slower than the other card obviously. Until today.

Today, I decided to load up The Sims 4, a game I have not played in a month. Since the last time I played, there was an update to the game, so I updated the game and the mods I've installed for the game. First sign that something was wrong was when I loaded up a save game and my Sim is gone. I exited the game and updated one more mod that I previously didn't. Then the really bad thing happened, I launched the game again, then bam, blackscreen and then restart, followed by a WHEA 18 Cache Hierarchy error in the Event Viewer afterwards. After processing in my mind what the hell just happened, I repaired The Sims 4 through the EA app, loaded up the game again, and this time around it did not crash, and I was able to load into the save game where my Sim was previously gone, and lo and behold my Sim is there. I played for a while, then exited the game and played Destiny 2 afterwards to make sure this problem doesn't occur again. It didn't. But hardly reassuring, given how intermittent my previous WHEA 18 errors were. So I want to ask, can a corrupt game outright cause a WHEA 18 error? Or was that just a coincidence and one of my hardware is in fact failing?

Things I have done previously:

• Updated graphics drivers to the latest, AMD Adrenalin 23.3.2.
• Overclocked the poor RX550 using MSI Afterburner, core to 1300Mhz from default 1206Mhz, memory to 1600Mhz from default 1500Mhz, power limit +50%, could this be the reason? I have since removed the overclock, I don't think it made much of a difference in games anyway.
• BIOS is up to date, version 4.60.
• So is the chipset driver.
• Loaded up Stable Diffusion in CPU only mode to see if my CPU is failing. Nope, just high-ish temps of around 75C but no crash.
• Tested memory with memtest86+ and Testmem5, no errors, not even with XMP on.
• My RX550 doesn't go above 66C with the overclock mentioned above. But that's just the GPU temps, the card doesn't have sensors to measure VRAM temps.
• CrystalDiskInfo shows that my NVME C: drive is at 98% health.
• My wifi card the TP-Link WN881ND does crash my PC sometimes, with either a Driver IRQL or System Thread Exception BSOD. I use the drivers from the TP-Link website, but that isn't the latest dated one.
• sfc /scannow says "Windows Resource Protection did not find any integrity violations."

 

[OneMoar]

cache hieracy is cpu core instablity

usually pbo or curve optimiser at fault

 

[3x0]

I've seen Cache Hierarchy error caused by both CPU and GPU VRAM instabilities. If the errors continue with GPU at stock clocks, then stress test the CPU with something like CoreCycler.

 

[RJARRRPCGP]

This seems common with unstable VRAM, strangely. (RAM on video card)

 

[A Computer Guy]

Have you tried bumping your SOC voltage up a bit (1.1v) or loosen your RAM timings to see if it gives some more stability?

 

[emothxughts]

Have you tried bumping your SOC voltage up a bit (1.1v) or loosen your RAM timings to see if it gives some more stability?

Here's my timings and SOC volts as reported by Zentimings:

I set the SOC volts to 1.15v in BIOS. I am not entirely familiar with the concept of loosening RAM timings, but these are the timings that my XMP profile gave me.

cache hieracy is cpu core instablity

usually pbo or curve optimiser at fault

I have done CPU stress tests with Cinebench R23, Prime95 and OCCT and I'm also running Stable Diffusion on CPU only mode while I'm writing this. No crashes, just high temps (but somewhat below 80C, which is way below 95C). If I recall correctly, PBO is set to Auto in BIOS and I never even touched Curve Optimizer.

 

[A Computer Guy]

Here's my timings and SOC volts as reported by Zentimings:
View attachment 289710
I set the SOC volts to 1.15v in BIOS. I am not entirely familiar with the concept of loosening RAM timings, but these are the timings that my XMP profile gave me.

First I would simply test one thing at a time. Rerun Sims and see if it helped or not.

To clarify was that (in the screenshot) the voltage before you changed it?

 

[emothxughts]

This seems common with unstable VRAM, strangely. (RAM on video card)

I sure hope this is just the result of my VRAM not taking too kindly to being OC'd to 1600MHz, that it wants to stay at stock 1500MHz and below.

 

[freeagent]

I had an amd card that would make the screen blink with just 1mhz OC on the vram.. that was an Asus card

 

[emothxughts]

First I would simply test one thing at a time. Rerun Sims and see if it helped or not.

I'll run the game again when I have the time. I did rerun the game right after repair though, and it played fine, for just 5 minutes though, was busy with something else.

Also that was what I (actually the XMP/DOCP profile) set the RAM to, long before this new WHEA 18 error happened.

 

[OneMoar]

whea 18 is cpu related there is no code 18 for anything related to the gpu. that would be tdr event or visable graphics corruption or dxgi bsod

folks don't suggest random stuff without verifying what the problem accually is please

check event log for which APIC failed.(should be in the details pane of the whea event) either your cpu is slighly degrading or pbo isn't as stable as you thought. or the motherboard defaults are not stable. baring that monitor voltages to see if maby the psu is out of spec

but this is 100% a cpu problem nothing else throws cache hierarchy errors except imc/cpu core
the fact that its stable at full load indicates something funky is happening at boost clocks this is usually voltage related

 

[A Computer Guy]

I'll run the game again when I have the time. I did rerun the game right after repair though, and it played fine, for just 5 minutes though, was busy with something else.

Also that was what I (actually the XMP/DOCP profile) set the RAM to, long before this new WHEA 18 error happened.

XMP/DOCP should setting the DRAM voltage, usually SOC is defaulted to AUTO as far as I am aware. If the SOC bump didn't help I'd set it back to AUTO.
If I recall correctly SOC helps with infinity fabric stability in the CPU so that is why I suggested it.

 

[3x0]

folks don't suggest random stuff without verifying what the problem accually is please

I've had that error on a system with unstable VRAM, talking from experience. Although, it might be only with ReBAR enabled.

 

[emothxughts]

whea 18 is cpu related there is no code 18 for anything related to the gpu. that would be tdr event or visable graphics corruption or dxgi bsod

folks don't suggest random stuff without verifying what the problem accually is please

check event log for which APIC failed.(should be in the details pane of the whea event) either your cpu is slighly degrading or pbo isn't as stable as you thought. or the motherboard defaults are not stable. baring that monitor voltages to see if maby the psu is out of spec

but this is 100% a cpu problem nothing else throws cache hierarchy errors except imc/cpu core
the fact that its stable at full load indicates something funky is happening at boost clocks this is usually voltage related

That's what I thought at first when I encountered my first ever WHEA 18, that my R5 3600 was defective. So I RMA'd it, put my previous R3 3200G back into the system running the 2nd hand RX5600XT card I also recently bought then, then bam, same crash & restarts happen. Later I got a replacement R5 3600, and the exact same crash & restarts happened. Read on Reddit that some WHEA 18 errors are caused by bad GPUs, so I deduced that my RX5600XT card was causing it all. So I sent it out for repair, put my emergency RX550 in, and I've been running games crash free ever since, at least before this Sims 4 situation.

On a side note, thank you TPU for your prompt response to this post. I posted this exact same problem on the LTT forums, but I might as well be talking to a wall, because nobody replied.